Regime AlphaCheck the context.
Confirm the analysis date, portfolio and any stale-input warning.
Start here / Your first review
From your first holding
to a clearer review.
Know where to start, what each tab offers, and how to reach evidence you can explain.
The example is public and synthetic. The private app requires an approved account. Requesting access does not create an account.
First-time path
Four steps. Three explicit Builds.
Bring tickers, capital weights and the period you want to investigate. Check supported history before choosing the analysis window.
- Step 1Enter your portfolio
Add tickers and capital weights totaling 100% in Overview. These are draft inputs.
Overview - Step 2Build the model
In Research, check Features and Model, then choose Build Model. Resolve any readiness issue before continuing.
Research - Step 3Build measured risk
In Portfolio risk, choose Build exposures, then Build risk analysis. Each stage uses explicit controls.
Portfolio risk - Step 4Review and keep the evidence
Return to Overview for the portfolio story. Use Review / export to keep a brief and restorable run bundle.
Review / export
Build blocked? Read the reason → correct the indicated input or coverage → retry the explicit Build. Earlier successful results stay available.
Before choosing variables or dates
Data & Variables is an optional preparation area. Refresh snapshot metadata, then use the refreshed snapshot for the next Build. Inspect source, units and coverage, select inputs, choose Build coverage preview, then Use these inputs in Research. Feature inputs and portfolio-pricing history have different roles.
A supported ticker can still lack history for your chosen dates. Keep the holding's actual identity and correct the request explicitly; the app does not silently substitute an instrument or invent prices. Model and exposure settings determine what can be estimated, not whether an investment strategy works.
Returning-user path
Start with the evidence you saved.
Inspect the difference.
Read overall drivers, select a supported regime, then check changes and uncertainty.
Keep a useful record.
Export a brief and run bundle. Session history is temporary; restore a compatible bundle after reconnecting.
Changing a tab or chart does not refresh data or fit a model. Draft edits leave saved evidence intact until the relevant explicit Build succeeds.
The workspace at a glance
Which tab answers your question?
Open a tab below for its prerequisites, output and recommended next step. App links require sign-in.
OverviewWhat should I review first?
Core
Enter holdings; return here to read the saved portfolio story.
Requires: Holdings to start; completed Builds for results.
Produces: Draft weights, saved history, drivers and review questions.
Next: Research
Open Overview in the appStart hereWhere do I begin or go next?
Help
Follow the workflow and find the right tab.
Requires: An approved account; no analysis required.
Produces: A next step and a map of the workspace.
Next: Overview
Open Start here in the appResearchHow do I build the model?
Core
Use Portfolio → Features → Model → Build; press Build Model explicitly.
Requires: Valid weights, selected features and supported dates/history.
Produces: Saved portfolio history, fitted regimes and diagnostics.
Next: Portfolio risk
Open Research in the appData & VariablesWhich inputs have usable history?
Optional preparation
Inspect sources/units, connect published metadata, preview coverage and apply selected features to Research.
Requires: A published snapshot for file verification and previews.
Produces: A feature draft and coverage evidence; no fitted model.
Next: Research
Open Data & Variables in the appPortfolio riskWhat carries risk, overall and by regime?
Core
Build exposures, then Build risk analysis; inspect signed driver comparisons.
Requires: A successful Build Model; supported sleeve history for exposures.
Produces: Measured exposures, reconciled risk and conditional evidence.
Next: Overview
Open Portfolio risk in the appRegime evidenceHow are the estimated states supported?
Supporting review
Inspect saved probabilities, timing, sample support and diagnostics.
Requires: Saved model inference.
Produces: Evidence about this model's states and their limitations.
Next: Review / export
Open Regime evidence in the appScenariosHow do explicit alternative weights compare?
Optional analysis
Compare a declared proposal or benchmark using the saved covariance.
Requires: Compatible saved risk and measured history for every instrument.
Produces: Historical comparisons; no forward stress probabilities.
Next: Review / export
Open Scenarios in the appReview / exportHow do I keep and explain this analysis?
Core
Review/export a brief and owner-bound bundle, or restore a compatible bundle.
Requires: A saved run to export; a valid owner-bound bundle to restore.
Produces: Portable evidence and session history.
Next: Overview
Open Review / export in the appConditional PCA ResearchWhat statistical structure deserves deeper inspection?
Advanced research
Inspect declared matrix semantics and principal eigenspaces.
Requires: Supported saved exposure/risk evidence; explicit Builds for new work.
Produces: Statistical diagnostics, not causal economic drivers.
Next: Portfolio risk
Open Conditional PCA Research in the appRead a chart in three passes.
- Confirm the saved portfolio, dates and units.
- Read the overall result, then compare a supported regime.
- Inspect exact values, sample support and limitations before drawing a conclusion.
Regimes are model-defined states. Historical membership is not forecast confidence, and a negative risk contribution does not establish lasting diversification.
Want to use your own portfolio?
Email Joe about your use case. He can confirm availability and the account to use. No account is created by sending an email.
Request access